计算任何给定日期的观察值

Question

我正在使用类似于以下内容的数据集：

df <- data.frame(type = c("A", "A", "A", "A", "A", "B", "B", "B", "C", "D", "D", "D"), 
                 start_date = as.Date(c("2010-02-01", "2011-03-15", "2011-09-15", "2015-01-01", "2015-05-15", "2009-01-01", "2015-07-14", "2016-06-30", "2012-01-15", "2010-04-05", "2010-08-01", "2012-04-01"), format = "%Y-%m-%d"), 
                 end_date = as.Date(c("2010-12-31", "2011-07-31", "2014-04-04", "2015-02-15", "2016-12-15", "2013-02-16", "2015-12-31", "2016-12-31", "2015-09-17", "2010-04-10", "2010-09-30", "2013-12-31"), format = "%Y-%m-%d"))

我想计算任何给定日期的观察次数。

预期输出

从本质上讲，我希望我的结果表明只有一种type到 2010-02-01，然后两种到 2010-04-05，然后三种到 2010-04-10 等等，即一列带有日期（每行day) 和一列type .

date count_of_type
2009-01-01 1
2009-01-02 1
2009-01-03 1
...
2010-01-31 1
2010-02-01 2
2010-02-02 2
...
2010-04-04 2
2010-04-05 3
2010-04-06 3
2010-04-07 3
2010-04-08 3
2010-04-09 3
2010-04-10 2
2010-04-11 2
...

我认为这很容易做到，但无法弄清楚......知道吗？

干杯，

Answer 1

一种选择是transmute通过采取相应的seq由“起始日期”，“END_DATE”的uence day ，然后拿到count

library(tidyverse)
df %>% 
   transmute(date = map2(start_date, end_date, seq, by = '1 day')) %>% 
   unnest %>% 
   count(date)